Progress In Electromagnetics Research M, Vol.

4, 6780, 2008

SMART ANTENNA BASED DS-CDMA SYSTEM

DESIGN FOR THIRD GENERATION MOBILE
COMMUNICATION
A. Kundu
Kalpana Chawla Space Technology Cell (KCSTC)
Department of Electronics & Electrical Communication Engineering
Indian Institute of Technology
Kharagpur 721302, India
S. Ghosh
Netaji Subhash Engineering College (NSEC)
Department of ECE
West Bengal University of Technology
India
B. K. Sarkar and A. Chakrabarty
Kalpana Chawla Space Technology Cell (KCSTC)
Department of Electronics & Electrical Communication Engineering
Indian Institute of Technology
Kharagpur 721302, India

AbstractThe third generation promises increased bandwidth up

to 384 Kbits/s for wide area coverage up to 2 Mbits/s for local
area coverage. A smart antenna technique has been developed
specically to meet the bandwidth needs of 3 G [5, 7, 12, 14]. Smart
beamforming provides increased data throughput for mobile high
speed data application. Here in this paper we worked on MVDR
Beamforming scheme for continuing demand for increased bandwidth
& better quality of services. Adaptive minimum variance beamforming
can be easily implemented to increase capacity as well as suppressing
co-channel interference & to enhance the immunity to fading. The
computer simulation carried out in MATLAB platform shows the signal
processing technique optimally combines the components in such a way
that it maximizes array gain in the desired direction simultaneously
minimize it in the direction of interference [1, 3, 15, 16].

68

Kundu et al.

1. INTRODUCTION
The use of an antenna array adds an extra dimension and makes the
utilization of spatial diversity possible. This is due to the fact that the
interferers rarely have the same geographical location as the user, and
therefore they are spatially separated. Dierent from Omni-directional
antenna system and sectored antenna system, adaptive antenna array
system referred in Fig. 1, combine an antenna array and a digital signal
processor to receive and transmit signals in a directional manner. Thus
the beampattern at the base station can be adaptively changed. The
beamformer has to satisfy two requirements. 1. Steering capability
whereby the desired signal is always having found the maximum gain.
2. The eect of sources of interferences is minimized [4, 68]. An
improvement in the system capacity in a multipath environment can
be achieved by combining a set of beamformer & a RAKE combiner.

Figure 1. Adaptive antenna with adaptive processor.

There are two minimum variance beamformer congurations for
the base station receiver in direct sequence code division access
(DSCDMA) system depending upon the chip & symbol conguration.
In symbol based conguration, spatial diversity is exploited after
dispreading & interferer components are rejected, based on both their
power and code correlation with the signal of interest (SOI). The
symbol based conguration is very ecient method to reject higher
interfering strength components. Fig. 2 shows the adaptive antenna
array with adaptive processor which focuses all transmitted power
to the user and only looks in the direction of the received signal.
This ensures that the user receives the optimum quality of services

Progress In Electromagnetics Research M, Vol. 4, 2008

69

Figure 2. Beamformer & DOA estimation.

and maximum coverage for a base station. Adaptive antenna array
incorporate more intelligence into their control system to estimate
the signal of choice. Adaptive antenna monitor their environment
and in particular, the response of the data path between the user
and the base station [10, 11, 14, 15]. This information is then used
to adjust the gains of the data received or transmitted from the array
to maximize the response of the user. This dynamic adaptation of
the antenna array response provides focused beams to specic users
and a new mechanism for multi-user access to the base station. The
fundamental operation carried out in adaptive arrays is to pass the
data stream from each antenna through an adaptive nite impulse
response lter. As per example when we have four antenna elements
in a narrow band environment such that the adaptive lter results in
a single multiplication, we realize processing requirements approach
one-half billion multiple accumulates per second (MACs) for a sample
rate of 105 mega sample per second. This sample rate is for a single
beam doesnt include the processing requirement for adaptive update
algorithm, so in this case we want to support multiple beams & achieve
ner beams by increasing the number of antennas, we could quickly
exhaust the processing capability of a standard processor architecture
as we reach processing requirements of several billion MACs per
second.
By using FPGAs, we have powerful DSP devices for handling
these high performance requirements at sampled data rates, further

70

Kundu et al.

more we can take advantage of the FPGA exibility for directly

handling acquisition control and other DSP function such as digital
down conversion, demodulation & matched ltering.
2. SMART ARRAY CONFIGURATION
Smart antenna implemented in the IMT-2000 mobile telecommunication systems to achieve a high system capacity. The following problems
become critical when expanding the use of W-CDMA in the telecommunication.
1) Interference in low transmission rate signals caused by high
transmission rate signals.
The number of users who use low
transmission rates while high transmission rate communication is in
progress will decrease because the high transmission rate signals will
cause a large amount of interference in the low transmission rate
signals.
2) There will be stronger demands to increase the number of users
a base station (BS) can accommodate. The number of users a BS
can accommodate is limited by the mutual interference from the users
signals and other factors [9]. Moreover, the area of a W-CDMA cell
is smaller than that of a second-generation mobile communications
system due to the restriction of transmission power. Therefore, to
expand the use of W-CDMA, it will be necessary to increase the
number of users that a base station can accommodate and also increase
the cell area. To solve these problems, we developed a smart sensor
array technology for a W-CDMA base station that reduces the amount
of interference that is received [2, 3, 57]. Smart array makes it possible
to increase the number of users. It also makes it possible to cover the
same area with a smaller number of base stations.
3. THE ALGORITHMS FOR SMART ANTENNA TEST
BED
The performance simulation relied on theoretical analysis depending
upon mathematical model and assumption. The construction of
the test bed is designed into two parts of DOA estimation and
beamforming.
3.1. MVDR Beamforming Scheme for Smart Antenna Test
Bed
Frequency domain snapshot consists of signal & noise components;
X() = Xs () + N (). Signal vectors can be represented by

Progress In Electromagnetics Research M, Vol. 4, 2008

71

Xs () = F ()v, where v is array manifold vector and F () represents

the frequency domain snapshots; the noise snapshot N () is a zero
mean random vector with spectral matrix Sn () = Sc () + 2 I. We
process X() with a matrix operation W H (). The dimension of
W H () is 1 N . The rst criterion of interest is called distortionless criterion. In the absence of noise Y () = F (), i.e., we wish to
minimize the variance of Y () in the presence of noise. We represent
Y () = F () + Yn () and minimize E[|Yn ()|2 ]. The constraint of
no distortion implies W H () v = 1. The mean square of output
noise is E[|Yn |2 ] = W H ()Sn ()W (). In this case we have taken
the minimum variance unbiased estimate E[Y ()] = F (). Now we
would impose a Lagrange multiplier




F
W H()Sn()W ()+() W H() v1 +() v H W ()1
(1)
The complex gradient with respect to W H () and solving we have
W0H () = ()v H Sn1 ()

(2)

() = [v H Sn1 ()v]1 ; W0H () = (v H )Sn1 ()

(3)

where

And = [v H Sn1 v]1 . The output of the optimum distortion less

matrix is the maximum likelihood estimate of F (m ) and can be used
to generate f (t)|ml . The MVDR lter provides the ML estimate of
signal f (t) when signal wave number is known and Sn () is known.
Hence
H

W0H () = ()v H Sn1 ()
WMVDR

(4)

The array gain at a particular frequency is given by the ratio of signal

spectrum to noise spectrum at the o/p of the distortion-fewer lters
compared to the ratio of the input. The input SNR at each sensor is
Sf ()
Sn () and array gain comes
Ac =

N2
v H n ()v

(5)

Here n () is the normalized spectral matrix. In the side lobe region

the MVDR Beamformer oers signicant improvement for large INR
and gives acceptable beam pattern. In the outer side lobe region
the beam pattern of MVDR beam-former starts to degenerate, the

72

Kundu et al.

HPBW region, the MVDR beam-former generates a solution that

distorts the main lobe and is sensitive to model mismatch. Fig. 3
shows a single channel of a smart array system which consists of
various RF components connected with a dedicated FPGA processor.

Figure 3. One channel of the smart antenna.

Progress In Electromagnetics Research M, Vol. 4, 2008

73

20

dB

10

Beam pattern (dB)

-10
-1

-0.8

-0.6

-0.4

-0.2

0.2

0.4

0.6

0.8

-0.8

-0.6

-0.4

-0.2

0.2

0.4

0.6

0.8

0
-10
-20
-30
-40
-1

/
Figure 4. Optimum beampattern autoregressive interference in the
absence of white noise when = 0.6.
Fig. 4 illustrates the case when multiple plane waves desired signals
are present. Array generates multiple beams to maintain the link.
Fig. 5 indicates normalized array gain of MVDR Beamformer in the
presence of mismatch when single plane wave interferer is present with
INR=10 dB.
3.2. DOA estimation:
MUSIC algorithm estimate direction-of-arrival (DOA) of a user.
However, the MUSIC spatial spectrum does not estimate the signal
power associated with each arrival angle and MUSIC fails when
impinging signals are highly correlated. We present max Eigen
value algorithm which has performances of the peak of spectrum
corresponding directly with main signal, good ability of resisting
multipaths and lower computational complexity. Fig. 6 shows the array
estimated the desired signals coming from 30 & 60 respectively.

74

Kundu et al.
1

A m vdr /A o ( v m )

0.9

0.8

0.7

0.6

0.5

0.4
-0.1

-0.08 -0.06 -0.04 -0.02

0.02

0.04

0.06

0.08

0.1

ua

Figure 5. Array gain of beamformer (MVDR).

4. MVDR FOR MULTIPLE PLANE WAVE
INTERFERERS
Here D dierent paths are considered which are normalized to steering
and sum devices corresponding to the D interfering noises. The
transform of the interfering noise can be written as
Nc () =

D

i=1

Ni ()v

and Snc () =

D
D 


Sni nj ()vv H

(6)

i=1 j=1

which is a N D dimensional matrix and VI = [v(k1 )v(k2 ) . . . v(kD )].



S11 () . . . S21 () . . .
...
; and
So the composite array manifold SI =
. . . SDD ()
this is the interference spectral matrix. We assume that the signal and
interference are statistically independent.
Sn = 2 I + VI SI VIH

(7)

Progress In Electromagnetics Research M, Vol. 4, 2008

75

The Sp at ial MUSI C Spectr um

0

The Angular Pseudo S pectr um

-5

-10

-15

-20

-25

-30

-35

20

40

60

80

100

120

140

160

180

Th e Phase Angle In Degrees

Figure 6. Plot of angular pseudo spectrum with phase angle for

spatial MUSIC when the channel signal to noise ratio: 20 dB & the
signal arrival phase angle: 30, 60 degree.
Using the matrix inversion lemna we have;

1
1
1
1
1
SI VIH 2
Sn1 = 2 I 2 VI I + SI VIH 2 VI

We have
W0H = vsH Sn1 =

N VsH
H

SI oc
2
N

(8)

(9)

where
Hoc

1
SI H
SI H
= I + 2 vI vI
v

2 I

(10)

this is a D N dimensional matrix lter whose output is the minimum

mean square error estimate nI (t) in the absence of desired signal,
H
SI = vsNvI a 1 D dimensional spatial correlation matrix. The

76

Kundu et al.

optimum receiver is estimating the directional noise vectors and then

subtracting the component of the estimate that is corrected with the
desired signal.
5. MVDR FOR MULTIPLE PLANE-WAVE SIGNALS
We have already introduced X() = vs F () + N () where vs =


F1 ()

are the Fourier

[v( : k1 ) . . . v( : kD )] and F () =
FD ()
transform of an unknown non random signal vector. The processor
W H () is a D N distortion less matrix processor W H () =

W1H ()

; suppressing the dependence the distortion less criterion
WDH ()
implies W1H vF = F1 ; W2H vF = F2 ; and WDH vF = FD . These set of
equations must be satised for arbitrary F , it places D constrains on
each Wi . This implied W1H v1 = 1; W1H v2 = 0 and simultaneously
W1H vD = 0. For ith beamformer W1H vj = ij where . . . , i, j =
1, . . . , D. Therefore, the ith beamformer is distortion less with respect
to the ith signal and put a perfect null of the other D-1 signals. The
2 = W H S W . We would try to
output noise of the ith beamformer ni
n i
i
2 subject to D constrainon W .
minimize ni
i
Dene,
Gi =

WiH Sn Wi

D


ij

WiH vi

ij +

j=1

D




ij viH wi ij

(11)

j=1

where i = 1, . . . , D minimizing we have,

WiH Sn + ij vjH = 0
or,

WiH

ij vjH Sn1

(12)
(13)

Now we dene a D D matrix whose i, jth element is ij . =

[V H Sn1 V ]1 and nally the optimum distortion less beamformer is

1 H 1
H
= V H Sn1 V
V Sn
Wd0

(14)

Figure 7 shows Comparison of MVDR, conventional and Null

steering beam pattern taking the number of antenna elements 10.
MVDR maximizes the output SNR does not require knowledge of
directions and power levels of interferences as well as the level of the

Progress In Electromagnetics Research M, Vol. 4, 2008

77

Interference Power 10 dB
20
Conventional
MVDR
NUll Steering

10

Beam pattern (dB)

-10

-20

-30

-40

-50
-1

-0.8

-0.6

-0.4

-0.2

0.2

0.4

0.6

0.8

Figure 7. Comparison of MVDR, conventional and null steering beam

pattern with N = 10.
20

Sensitivity (noise power) (dB)

15
10
5
0
-5
-10
-15
-20

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Figure 8. Sensitivity function (inverse of noise gain) vs ui for linear

array of 10 elements.

78

Kundu et al.

background noise power to maximize the output SNR. Fig. 8 shows

the variation of sensitivity for MVDR, Null steering & conventional
beamformer. We want the array processor to be very sensitive
to directional constraints, whenever it nds slight changes in the
environment it should adjust the weights accordingly, the algorithm
may be summarized in the ow chart of the Fig. 9.

Figure 9. Software ow in FPGA module for adaptation.

6. CONCLUSION
We have simulated a multipath environment, taken that the mobile
unit move at a distance of 50 to 100 meter away from test bed and
round by center of test bed. Here we have taken two user, user1 as
example with 0 dB power at angle 30 (DOA) and user2 with 6 dB of
power and having DOA 60 respectively. By our smart antenna test
bed at experimental locale it can be seen the MUSIC algorithm can
estimate consistently with desirable result. After estimation of DOAs
of dierent signals, may be fed to beamforming network. Fig. 5 shows
the MVDR Beamformer average gain. Two simulated beampattern of
antenna array indicate that the array is capable of steering the beam
in the direction of the users whereby creates null in the direction of
interferer.

Progress In Electromagnetics Research M, Vol. 4, 2008

79

REFERENCES
1. Godara, L. C., Error analysis of the optimal antenna array
processors, IEEE Trans. Aerosp. Electron. Syst., Vol. 22, 395
409, 1986.
2. Suard, B., A. F. Naguib, G. Xu, and A. Paulraj, Performance
of CDMA mobile communication systems using antenna arrays,
IEEE Int. Conf. Acoustics, Speech, and Signal Processing
(ICASSP), 153156, Minneapolis, MN, 1993.
3. Gebauer, T. and H. G. Gockler, Channel-individual adaptive
beamforming for mobile satellite communications, IEEE J.
Select. Areas Commun., Vol. 13, 439448, 1995.
4. Chen, C.-H., C.-L. Liu, C.-C. Chiu, and T.-M. Hu, Ultra-wide
band channel calculation by SBR/image techniques for indoor
communication, J. of Electromagn. Waves and Appl., Vol. 20,
No. 1, 4151, 2006.
5. Fakoukakis, F. E., S. G. Diamantis, A. P. Orfanides, and
G. A. Kyriacou, Development of an adaptive and a switched
beam smart antenna system for wireless communications, J. of
Electromagn. Waves and Appl., Vol. 20, No. 3, 399408, 2006.
6. Babayigit, B., A. Akdagli, and Kerim Guney, A clonal selection
algorithm for null synthesizing of linear antenna arrays by
amplitude control, J. of Electromagn. Waves and Appl., Vol. 20,
No. 8, 10071020, 2006.
7. Vescovo, R., Beam scanning with null and excitation constraints
for linear arrays of antennas, J. of Electromagn. Waves and
Appl., Vol. 21, No. 2, 267277, 2007.
8. Noori, N. and H. Oraizi, Evaluation of MIMO channel capacity
in indoor environments using vector parabolic equation method,
Progress In Electromagnetics Research B, Vol. 4, 1325, 2008.
9. Koo, B.-W., M.-S. Baek, and H.-K. Song, Multiple antenna
transmissiontechnique for UWB system, Progress In Electromagnetics Research Letters, Vol. 2, 177185, 2008.
10. Fallahi, R. and M. Roshandel, Eect of mutual coupling and
conguration of concentric circular array antenna on the signalto-interference performance in CDMA systems, Progress In
Electromagnetics Research, PIER 76, 427447, 2007.
11. Mahmoud, K. R., M. El-Adawy, S. M. M. Ibrahem, R. Bansal, and
S. H. Zainud-Deen, A comparison between circular and hexagonal
array geometries for smart antenna systems using particle swarm
optimization algorithm, Progress In Electromagnetics Research,
PIER 72, 7590, 2007.

80

Kundu et al.

12. Gozasht, F., G. Dadashzadeh, and S. Nikmhr, A comprehensive

performance study of circular and hexagonal array geometreis in
the LMS algorithem for smart antenna applications, Progress In
Electromagnetics Research, PIER 68, 281296, 2007.
13. Mukhopadhyay, M., B. K. Sarkar, and A. Chakrabarty, Augmentation of anti-jam GPS system using smart antenna with a
simple DOA estimation algorithm, Progress In Electromagnetics
Research, PIER 67, 231249, 2007.
14. Mouhamadou, M., P. Vaudon, and M. Rammal, Smart
antenna array patterns synthesis: Null steering and multi-user
beamforming by phase control, Progress In Electromagnetics
Research, PIER 60, 95106, 2006.
15. Varlamos, P. K. and C. N. Capsalis, Electronic beam steering
using switched parasitic smart antenna arrays, Progress In
Electromagnetics Research, PIER 36, 101119, 2002.
16. Sanyal, S. K., Q. M. Alfred, and T. Chakravarty, A novel beamswitching algorithm for programmable phased array antenna,
Progress In Electromagnetics Research, PIER 60, 187196, 2006.